Email: dbmster (at) gmail
Education
PhD, Economics, UCLA, 2012
MS, Financial Mathematics, Stanford University, 2005
BA, Economics, Mathematics, Stanford University, 2004
Papers
Identification of Linear Regressions with Errors in all Variables, forthcoming in Econometric Theory
We provide necessary and sufficient conditions for identification of the linear errors-in-variables model, extending Reiersol (1950) from the single regressor model to the multiple regressor model. We introduce a new estimator for the coefficients using second derivatives of the empirical log characteristic function.
Identification of Joint Distributions in Dependent Factor Models (2018), Econometric Theory
We identify and estimate joint distributions in a factor model with subsets of arbitrarily dependent unobservables. Identification extends a result by Kotlarski (1967) from a single factor model to a multiple factor model.
Identification of Additive and Polynomial Models of Mismeasured Regressors Without Instruments (2017), with Xavier D’Haultfœuille and Arthur Lewbel, Journal of Econometrics
We identify the nonparametric function g in the model Y = g(X*) + h(Z) + U, E[U|X*,Z]=0 where X* is measured with error and without any side information. Similarly, we identify the polynomial P in the model Y=P(X*,Z) + U, E[U|X*,Z]=0 .
Identification of Linear Factor Models with Subsets of Arbitrarily Dependent Factors
We extend identification from the factor model with mutually independent factors and the errors-in-variables model with arbitrarily dependent unobserved regressors to factor models with subsets of arbitrarily dependent factors.
Stochastic Frontier Estimation of Housing Supply, with David Genesove
We estimate supply and cost of multi-level housing in unregulated markets, using only observed equilibrium prices and quantities. We find (1) economies of scale at low heights, relatively flat marginal costs for a range of heights, and steeply increasing marginal costs at high heights, (2) elasticity of substitution of land for capital is less than 0.4 so that it is hard to build up, (3) regulatory tax is about 50% of observed market prices (with higher tax in more expensive and denser areas) so that suppliers in unregulated markets would still find it profitable to build tall buildings, and (4) the effects of measurement error and quality differences on price are small relative to regulatory tax.